Medical Image Classification on Skin
51.91Top-1 AccuracyGIF
Evaluation results
| Method | Links | |
|---|---|---|
| GIFBackbone=ResNet-50, Training Protocol=Trained from scratch, Diffusion Model=Fine-tuned, Number of Samples=502026.03 | 51.91 | |
| Real GuidanceBackbone=ResNet-50, Training Protocol=Trained from scratch, Number of Samples=502026.03 | 51.81 | |
| Diff-AugBackbone=ResNet-50, Training Protocol=Trained from scratch, Number of Samples=502026.03 | 51.13 | |
| BaselineBackbone=ResNet-50, Training Protocol=Trained from scratch, Number of Samples=502026.03 | 50.48 | |
| CutMixBackbone=ResNet-50, Training Protocol=Trained from scratch, Number of Samples=502026.03 | 49.05 | |
| DA-FusionBackbone=ResNet-50, Training Protocol=Trained from scratch, Number of Samples=502026.03 | 49 | |
| Diff-MixBackbone=ResNet-50, Training Protocol=Trained from scratch, Number of Samples=502026.03 | 47.85 | |
| MixupBackbone=ResNet-50, Training Protocol=Trained from scratch, Number of Samples=502026.03 | 47.14 | |
| Real GuidanceBackbone=ResNet-50, Training Protocol=Trained from scratch, Number of Samples=102026.03 | 42.29 | |
| GIFBackbone=ResNet-50, Training Protocol=Trained from scratch, Diffusion Model=Fine-tuned, Number of Samples=102026.03 | 40.95 | |
| MixupBackbone=ResNet-50, Training Protocol=Trained from scratch, Number of Samples=102026.03 | 40.67 | |
| BaselineBackbone=ResNet-50, Training Protocol=Trained from scratch, Number of Samples=102026.03 | 39.62 | |
| Diff-AugBackbone=ResNet-50, Training Protocol=Trained from scratch, Number of Samples=102026.03 | 39.03 | |
| DA-FusionBackbone=ResNet-50, Training Protocol=Trained from scratch, Number of Samples=102026.03 | 38.71 | |
| Diff-MixBackbone=ResNet-50, Training Protocol=Trained from scratch, Number of Samples=102026.03 | 37.92 | |
| CutMixBackbone=ResNet-50, Training Protocol=Trained from scratch, Number of Samples=102026.03 | 36.86 |